This application claims priority to provisional application Ser. No. 60/091,431, filed Jul. 1, 1998 and Ser. No. 60/133,435, filed May 11, 1999, which both have the same title as this application and which are both incorporated herein by reference.
One of the challenges in the development of therapeutic compounds is to find a small molecule that is able to mediate a desired biological effect. Traditionally, synthetic chemistry and natural product screening have been the principal means for the derivation of many drug products.
High-throughput random screening is a standard procedure adopted by pharmaceutical companies for the discovery of lead compounds. This method relies upon availability of a large chemical database of natural/medicinal products. This procedure does not require the knowledge of principle components of biomolecules that cause disease. In short, it is a blind process to screen therapeutic lead compounds. The advantage of this approach is that it facilitates to build a large medicinal chemical database and can be repeatedly used to screen therapeutic compounds. Unfortunately, random screening is tedious and often requires isolation and characterization from natural extracts. Natural products are complex and include the stereochemical complexities inherent in their natural origin.
While high-throughput random screening procedures have been used to identify some novel therapeutic molecules, such procedures are often limited by the availability of large chemical databases. Advances in computer technology and the understanding of protein-protein interactions has allowed for attempts to replace the high-throughput screening procedures with computer-aided analysis and design of novel molecules. Such structure based approaches have reduced the time and resources to discover novel compounds.
Structure based approaches have been used to develop several inhibitors that are either “substrate analogs” or “allosteric” inhibitors. Allosteric effectors, in some cases, are considered superior to conventional substrate analog for reasons: (1) it is non-competitive with natural ligand, (2) it can be effective at a lower concentration, (3) allosteric binding sites are less conserved and thereby specificity and selectivity can be enhanced, and (4) in some cases allosteric effectors can inhibit the target molecules' function by trapping it in an intermediate non-native or molten globular state.
Structure-based approaches represent a targeted pathway where therapeutic agents are designed towards the biomolecule responsible for disease. There are two major approaches in the design of lead therapeutic compounds based on the nature of the molecule. For enzymes, design of substrate analogs (from the knowledge of active site) and peptidomimetics that has shown promise in some cases.
Substrate analogs are developed to compete with the natural substrate and occupy the active site. Thus, a potent therapeutic compound must have high affinity, exhibit selectivity and have longer retention time. Substrate analogs are better suited for enzymes, because many receptors and other non-enzyme molecules, such as receptors and their ligands have no defined active site but alter biological function. In such cases, a peptide's ability to mimic a protein's local structural features is one of the ways used to design therapeutic compounds. Substrate analog interactions are often not reversible.
Peptidomimetics are developed both as therapeutic agents and as a probe to understand biological functions. Natural products targeting opioid and hormone receptors are historical examples of peptidomimetics because they validate many of the concepts invoked in rational design. These compounds provide a classic example of how structurally different non-peptides may be from their peptide parents (lacking flexibility, amide bonds and obvious pharmacophore similarity) and how their modification can lead to highly selective ligands for subtypes of receptors for both peptide and non-peptide compounds.
Elucidation of the conformation of a peptide can provide insights about the structural requirements of its binding to a receptor (Boteju, L. W. et al., 1996, J. Med. Chem., 39:4120-4; and Cho, M. J. et al., 1996, Trends in Biotechnology, 14:153-8; which are both incorporated herein by reference). A major problem, however, in structure-activity studies of linear peptides is the large degree of flexibility, not only of the side-chain residues, but also of the peptide backbone. Substitution of individual amino acids followed by biological screening might reflect affective differences on structure rather than on residues implicated in binding. Consequently, spectroscopic studies in solution, where a rapid equilibrium between numerous conformations is likely to occur, have had little impact on the design of linear peptide analogues. In contrast, constrained peptides delineate solution conformations for correlation with receptor bound conformations. Bioactive compound design based upon conformational constrained peptide analogs representative of the recognition elements of the protein constitutes an effective approach to mimetic drug design. Constraints imposed upon peptides to lock in a particular conformation often times emulate those imposed by the tertiary structure of protein ligands. Imposed constraints can reflect the use of amino acids that contribute to the propensity of a particular secondary structure such as amphipathic helical repeats.
Despite the diverse usefulness of peptidomimetics, they remain less viable drugs due to their poor bioviability. Nevertheless, active peptide analogues with modified bonds or side chains, provide another approach in defining bioactive conformations and are valuable pharmacological probes, because generally they are more resistant to proteolytic degradation.
Protein structures have been elucidated using crystallography, NMR and molecular modeling. The three dimensional structures of proteins reveal (1) overall folding of the molecule, (2) scaffolds: secondary structural features such as α-helix, β-sheet, (3) functional units; b-turns and loops, and (4) surfaces that include cavities, clefts, pockets and crevices formed by the folding of amino acid chains on itself and, in the case of multimeric protein complexes, on itself and the amino acid chains of other subunits. Cavities, clefts, pockets and crevices can accommodate water molecules within an interior. Depending upon the nature of the amino acids which form the cavities, clefts, pockets and crevices molecules, the interior of these structural features have specific chemical and electrostatic properties as well as spatial dimensions.
Determination of crystal structures of proteins/receptors have provided a basic understanding of protein/receptors' function. Several receptors such as EGF receptors are activated either by ligands or by association with other erbB family of receptors. One of the hypothesis is that conformational changes induced either by ligand or by co-receptors elicits signal transduction. Thus, it is presumed that through allosteric mechanisms receptors can modulate signal transduction. Allosterically driven biological functions are also known both in enzymes and receptors (Ellis, J., 1997, Drug. Dev. Res., 40:193-204, and Kundrot, C. E. et al., 1991, Biochem., 30:1478-1484, which are both incorporated herein by reference). Attempts to modulate the function of proteins/receptors have been made and often referred to as “allosteric modification or allosteric inhibitors”.
Allosteric modification is a well known technique that has been studied in several enzymes (Iverson, L. F. et al., 1997, Protein Science, 6:971-982; Ladjimi, M. M. et al., 1985, J. Mol. Biol., 186:715-724; Ozaita, A. et al., 1997, Brit. J. Pharm., 121:901-912; Tang, J. et al., 1997, Chemistry & Biol., 4:453-459; and Tijane, M. et al., 1989, FEBS Lett., 245:30-34; which are each incorporated herein by reference) and receptors (Berthold, M. et al., 1997, Neurochem. Res., 22:1023-1031; Ellis, J., 1997, Drug. Dev. Res., 40:193-204; Kolliasbaker, C. A. et al., 1997, J. Pharmco. Exp. Therap., 281:761-768; and Robichon, R. et al., 1997, Eur. J. Pharmco., 328:255-263 which are each incorporated herein by reference). Hitherto techniques often used mutagenesis or small molecules identified from screening. Allosteric modifications have been used in enzymes to alter the enzymes' kinetics and in some cases used to develop inhibitors.
There is a need for modulators of intermolecular interactions and for methods of identifying such modulators. There is a need for inhibitors of intermolecular interactions and for methods of identifying such inhibitors. There is a need for enhancers of intermolecular interactions and for methods of identifying such enhancers. Structure based ligand design, as practiced today, requires the knowledge of cavity of known functions such as active sites, or cavities identified by high throughput (ligand binding). There is a need for a generalized approach to identify functional cavities for novel ligand design.